VEXYL Studio

The AI Agent Studio
for Voice and Chat

Build your AI agent once, then let it answer calls, make calls and reply to chats. Connect it to your SIP trunk, Asterisk or the PBX you already own, and run it self-hosted or on VEXYL Cloud.

One agent. Every conversation.

The same agent, with the same prompt, tools and knowledge, works across all three channels.

Receive calls

Answer every inbound call from your DIDs, IVR or PBX extensions. Understand the caller, look things up, book, and transfer to a human with context when needed.

Make calls

Place outbound calls through the API with per-call context, or run reminder, collection and survey campaigns from a CSV with the AI auto-dialer: up to 50 calls at once, retries and result webhooks.

Text chat

Put the same agent on your website, app or messaging integration through the REST and WebSocket chat API, with streaming replies.

See the AI auto-dialer for outbound campaigns

Build, connect, deploy

01

Build the agent

  • Prompt, variables and model per agent
  • Tools to book, look up, transfer or hang up
  • Knowledge base from your documents
  • Visual flows for scripted paths
02

Connect your lines

  • SIP trunk from your carrier
  • Asterisk and FreePBX
  • Panasonic, Avaya, Grandstream, Bitvoice PBXs
  • WebRTC widget for your website
03

Deploy and improve

  • Self-hosted or VEXYL Cloud
  • Every conversation recorded with cost
  • Analytics, traces and evidence
  • Webhooks to your CRM or workflow

Works with the phone system you already have

No rip-and-replace. Add AI agents as a SIP trunk or as a PBX extension.

SIP trunks

Carrier trunks with or without registration; G.711 and Opus.

Asterisk & FreePBX

AudioSocket for lowest latency, or a plain SIP trunk.

Business PBXs

Panasonic KX-NS / KX-NSX, Avaya IP Office, Grandstream UCM, Bitvoice.

Web & apps

WebRTC browser widget and chat API for text conversations.

PBX connection guides

Under the hood

VEXYL Studio brings together two VEXYL products. Use them together, or each on its own.

₹0.83 / min

All-in AI cost on 156 live Malayalam patient calls at a Kerala hospital, February 2026.

See the breakdown

Run VEXYL Studio your way

Same platform, two ways to run it. Start self-hosted for free, or let us run it for you on VEXYL Cloud.

Self-hosted

Run the gateway and LiteBot on your own servers or VPC. Your audio, transcripts and keys never leave your infrastructure.

  • Free Developer edition, 4 concurrent calls
  • Single binary or Docker
  • Bring your own SIP trunk and model keys
Self-host free

VEXYL Cloud

We run the gateway for you. Point a SIP trunk from your PBX at us, drop the browser widget on your site, and receive signed webhooks.

  • No servers or config files to manage
  • DID-to-bot mapping done for you
  • Opus, mu-law and A-law trunks
Talk to sales

Frequently Asked Questions

What is VEXYL Studio?

VEXYL Studio is an AI agent studio for voice and chat. You build an agent once - its prompt, model, tools and knowledge - and deploy it to receive phone calls, make phone calls and answer text chat. It connects to your SIP trunk, Asterisk or FreePBX, or a business PBX, and runs self-hosted or on VEXYL Cloud.

Can a VEXYL Studio agent both make and receive calls?

Yes. Agents answer inbound calls routed from your SIP trunk or PBX, and place outbound calls through the API or CSV campaigns with retries and result webhooks.

Which phone systems does VEXYL Studio connect to?

SIP trunks from carriers, Asterisk and FreePBX through AudioSocket or SIP, and IP PBXs including Panasonic KX-NS / KX-NSX, Avaya IP Office, Grandstream UCM and Bitvoice. Callers on your website can also talk to agents through the WebRTC browser widget.

Does the same agent handle chat and voice?

Yes. The same agent, with the same prompt, tools and knowledge base, can answer phone calls and text chat over the REST and WebSocket chat API.

Is VEXYL Studio self-hosted or cloud?

Both. Run it on your own servers for full data control, or use VEXYL Cloud and let us operate it for you.

Which AI models can I use?

Bring your own keys. Language models include Groq, OpenAI, Gemini, Vertex AI, OpenRouter, Sarvam and Ollama; speech providers include Sarvam, Google, Deepgram, ElevenLabs, Azure, Smallest and the open-source VEXYL-STT.